The collaborative use of data across value chains, e.g., through digital twins, holds significant economic potential. Estimates suggest that full adoption of digital twins in manufacturing could generate up to USD 37.9 billion annually. This potential is often hindered by connectivity issues due to the heterogeneity of technical systems and privacy concerns stemming from the perceived risk of losing control over sensitive information and intellectual property. These concerns limit inter-organizational data sharing, which is essential for the deployment of digital twins. This paper addresses the privacy challenge by conducting a requirements engineering for industrial data marketplaces based on academic literature. The resulting framework supports the design of platforms tailored to the needs of manufacturing stakeholders and aligns these needs with key principles from frameworks such as Gaia-X. To illustrate its relevance, a use case from fineblanking is presented, in which upstream data from material suppliers is used to enhance process monitoring via vertical federated learning to enable cross-organizational model training while preserving data privacy.
Mayer et al. (Thu,) studied this question.